빅데이터 분석 기반 개인 맞춤형 반영구 눈썹 디자인 자동화 모델 개발 = Development of an Automated Model for Personalized Semi- Permanent Eyebrow Design Based on Big Data Analysis

다국어 초록 (Multilingual Abstract)

This study aims to develop a big-data-based automated model for personalized semi- permanent eyebrow design by integrating facial shape information, landmark-based facial features, trend images, and expert design data. To address the limitations of conventional eyebrow design—primarily subjectivity and variability resulting from reliance on practitioner experience—this research quantitatively analyzes correlations between facial shape and key eyebrow design elements and incorporates these findings into the algorithm. A multi-stage recommendation system was constructed using decision trees, SVM, and Random Forest for facial shape classification and initial design selection, followed by CNN- and GAN-based generative modeling to produce natural and visually coherent eyebrow shapes resembling real treatment outcomes. The results show a facial shape classification accuracy of 92.7%, recommendation suitability of 85.9%, expert similarity of 86.4%, and a Harmony Index of 0.82, while user satisfaction reached 89%. Trend analysis revealed that a particular style accounted for 22% of the dataset and was incorporated as algorithmic weighting to reflect contemporary aesthetics and individual facial characteristics. Expert evaluation based on naturalness, harmony, professionalism, and procedural suitability was used as a validation metric, reinforcing the reliability of the generated designs. Overall, this study provides evidence supporting standardized, objective, and personalized semi-permanent eyebrow design and demonstrates the feasibility of automated design generation, contributing to AI-driven beauty services and the digital transformation of semi-permanent makeup practices.